Massimo Cristofanilli, MD, discusses the clinical utility of liquid biopsies—specifically focusing on circulating tumor cells and cell-free DNA—and highlights ongoing research in the space.
Massimo Cristofanilli, MD
In the pivotal SOLAR-1 trial, a liquid biopsy-based analysis of patients with breast cancer who harbored PIK3CA mutations was predictive of improved progression-free survival (PFS) outcomes with alpelisib, said Massimo Cristofanilli, MD. Cell-free DNA (cfDNA) was a major indicator of response to the drug, as it better reflected molecular status of the disease at the time of progression.
In the study, the combination of alpelisib and fulvestrant (Fulvestrant) nearly doubled median PFS compared with fulvestrant alone, according to data presented at the 2018 ESMO Congress.1 Median PFS in patients with PIK3CA mutations was 11.0 months for those who received the combination compared with 5.7 months for those who received placebo plus fulvestrant.
Cristofanilli, a professor of medicine at Northwestern University Feinberg School of Medicine, was a senior author on a study, findings of which supported the role of circulating tumor cells (CTCs) as a prognostic predictor. Overall survival (OS) data were taken from European cohorts and a group of patients from The University of Texas MD Anderson Cancer Center. Patients with advanced breast cancer were stratified by CTC enumeration taken from a liquid biopsy.
Additionally, plasma ctDNA samples were collected at baseline and analyzed by polymerase chain reaction to retrospectively assess PFS by PIK3CA mutation status, as a secondary endpoint of the SOLAR-1 trial. Results showed that the combination of alpelisib and fulvestrant led to a median PFS of 10.9 months versus 3.7 months with fulvestrant alone in those with PIK3CA mutation (HR, 0.55).2 In the non-mutant cohort, the median PFS was 8.8 months with the combination and 7.3 months with fulvestrant alone (HR, 0.80).
For all patients, CTC ≥5 was associated with a worse outcome (HR, 2.43; 95% CI, 2.17-2.73; P < .0001).3 Median OS for the indolent group (<5 CTC) was 36.3 months. In addition, for patients with de novo metastatic breast cancer prior to treatment, the indolent cohort demonstrated an OS greater than 5.5 years. This OS advantage for the indolent patients was maintained across all breast cancer subtypes represented in the study.
“The liquid biopsy is becoming more of a standard of care,” said Cristofanilli. “It is very clear we need to use it more and more, as soon as a patient is diagnosed with advanced breast cancer and there is the possibility to obtain a biopsy.”
In an interview with OncLive, Cristofanilli discussed the clinical utility of liquid biopsies—specifically focusing on CTCs and cfDNA—and highlighted ongoing research in the space.Cristofanilli: CTCs have been described as being associated with bad outcomes in patients with advanced breast cancer and other tumor types. For many years, this has been known. The clinical use has been very difficult because it has never been associated with any predictive information. We felt that because the prognostic value of CTCs in advanced breast cancer is [well] known, we wanted to have a definitive study confirming that CTCs are not only a bad outcome but [are indicative of] a different disease. We took data from large European cohorts—almost 2000 cases—plus approximately 500 cases taken from The University of Texas MD Anderson Cancer Center. We combined these data and asked the question, “Can we identify patients who have worse outcomes in the overall population and in the de novo established disease?” If that was the case, we do have different diseases, because that is the basis of staging.
We did exercise, as I said, almost 2500 cases. We are moving from calling this >5 or <5 CTCs to stage IV indolent and stage IV aggressive [disease]. We essentially had demonstrated [in a statistically significant fashion] that >5 is a much worse outcome in patients who have de novo disease and in the overall population. This is for first- and second-line settings, as well as later lines of therapy.
Now, this means a lot; it means that when we try to propose the standard of care a patient for their more aggressive or less aggressive disease, we have to consider them as different biologies. If we try to develop drugs in this space of metastatic disease, we have to at least stratify by CTC levels because these are 2 different diseases with different outcomes. It might take a much longer sample size of patients with indolent disease to show the difference.
Also, we tried to connect CTC-high and these more aggressive features with other [characteristics] of the liquid biopsy. We are looking into cfDNA and other mutations. One [set of findings] that we presented at the 2018 San Antonio Breast Cancer Symposium was that patients who had >5 CTC also had more mutations. This is all relevant, because it means that we not only have a prognostic signature, but molecular features that are actionable.Patients who have >5 CTCs have a worse outcome irrespective of the clinical features. The reason is that patients tend to progress in a way that is different than is seen in patients with less CTC. They develop more metastatic sites. The overall goal is to treat the metastases, and we need to identify these metastases and target them with therapies. If we do this effectively, we will finally be able to improve the survival of these patients.Liquid biopsies are a very complex field, but we can simplify it by looking at the cells, and cells can be enumerated, or we can look at the molecular features of the cell and cfDNA; this information is complementary, and we have to figure out how we can use it together.
One of the questions we ask is whether there are any features of cfDNA that we use right now, clinically, that we also gain from CTC. For example, if a patient has >5 CTC, what is the respective cfDNA [levels]? We saw it was associated with a high level of them, meaning it is not only a measure of tumor burden but of genomic instability. Here you have a patient with an aggressive disease and >5 CTC. When you plot all these, you have similar prognostic information. This has prognostic value, and it is not only predictive; you can use it for patients during treatment with the added value of molecular targets that we can treat properly.The liquid biopsy field is certainly moving forward very aggressively. The CTC field is moving more slowly and cfDNA needs to be analyzed in more phase III studies. We are seeing, for example, the SOLAR-1 study that was recently presented [that evaluated the use of liquid biopsies in patients with] PIK3CA mutations. The recent presentation identified cfDNA as a major predictor of response to the drug because it better reflected molecular status of the disease at the time of progression. In the context of the PALOMA-3 study, we are seeing the correlation of ESR1 mutations and PIK3CA mutations with treatment efficacy and with the dynamic of cfDNA being predictive of response to treatment and progression.It is very clear that when you use liquid biopsies, particularly cfDNA, in the context of the management of metastatic breast cancer, there are very actionable mutations. For example, we can target HER2 and PIK3CA mutations with directed therapy. ESR1 mutations are also making liquid biopsies more of a standard for advanced breast cancer. How we use this information to monitor patients and treat them is an unanswered question.